On Fri, Nov 25, 2011 at 23:17, Simon Horman <horms(a)verge.net.au> wrote:
I believe that the query-key option is what you are
after.
It should allow you to use the IP address that the client connect to
as one of the keys for the popmap lookup.
I did, eventually, see that, and felt silly afterwards for overlooking it. :)
If a user supplies a full email address (user@domain), I want to look
up the domain, and send their POP3 session to a specified server,
regardless of what IP they connected to. Failing that, I'll look up
the IP they connected to, and route their POP3 based on that.
For this, it looks like query_key should be \d,\I . Then, I assume
popmap looks like this:
domain1.com:10.1.1.1
domain2.com:10.1.1.2
192.168.1.1:10.1.1.1
192.168.1.2:10.1.1.2
(Fake IPs all around - the 10. IPs are the "real" back-end mail
servers, pretend the 192.168 IPs are public.)
I'm just having a bit of trouble wrapping my head round the fact that
there's only one popmap table, and I'll want it to do several things.
Thanks for the help!
David Smith
MVN.net